GeoDict Material Analysis Modules

With the GeoDict modules for Material Analysis, you can determine and compute the geometric and structural properties of a material. The analysis is based on 3D models — generated from 2D or 3D image data (µCT scans, FIB/SEM scans) or created with GeoDict's material modeling modules.

Depending on the material type, the analysis computes the proportion of solid materials (e.g. fibers, grains), binders, or pore space, and determines key parameters such as diameter, orientation, and volume.Artificial intelligence in the form of neural networks enables fast and precise characterization of even highly complex structures.

Combined with the Simulation & Prediction modules, the realistic 3D model can be used to simulate physical properties such as flow velocity, permeability, and effective elastic coefficients.A detailed understanding of material properties at the micro level helps you develop new materials and optimize production processes.

GeoDict Material Analysis includes four core modules: FiberFind, GrainFind, PoroDict, and MatDict.

1. FiberFind — AI-Powered Fiber Recognition

FiberFind represents the latest advancement in precise target recognition in µCT images, combining fast parameter prediction with neural network computation to accurately identify fibers and binders.It helps users understand 3D scan data of nonwovens, fiber-reinforced composites, and similar fiber-based materials.

After importing and segmenting CT or FIB/SEM scans, FiberFind offers three complementary analysis approaches:

Classic image processing to identify individual fibers

FiberFind-AI and BinderFind-AI for neural-network-based recognition

Statistical analysis of fiber diameter, orientation, and curvature distributions

Key Features

Fiber orientation distribution: Computes orientation tensors globally or per sub‑region (e.g., by layer); results can be exported to FiberGeo for structure reconstruction or to ElastoDict/ConductoDict for transversely isotropic material studies.

Curvature estimation: Generates statistical histograms of fiber curvature from µCT images.

Fiber diameter estimation: Calculates mean diameter and standard deviation; detailed diameter histograms show volume fractions by diameter for input into FiberGeo.

BinderFind-AI & FiberFind-AI: Neural-network-based tools to distinguish fibers from binders and identify individual fibers; GeoDict's unique structure generation (FiberGeo) provides reliable ground truth data for training.

Typical Applications

Material modeling in conjunction with FiberGeo

Binder content analysis in nonwovens and granular materials

Material optimization through virtual parameter studies before prototyping

Quality control by analyzing variations in fiber diameter, orientation, and curvature

FiberFind is particularly suited for fiber structures composed of long, solid, circular‑cross‑section fibers.Used together with FiberGeo, it forms a closed‑loop digital fiber material design workflow, reducing the need for physical lab experiments.

2. GrainFind — Grain and Binder Identification

GrainFind extracts statistical information from µCT‑scanned granular structures (e.g., battery electrodes).Its "Identify Grains" algorithm fits optimal shapes to each particle, identifying individual grains and their spatial orientation.The results can be exported to GrainGeo with one click to generate representative models of the microstructure.

In µCT images, binders and grains often share similar gray values and are difficult to distinguish. GrainFind addresses this with an AI‑based binder identification module that separates grains from binders based on shape differences.GeoDict includes pre‑trained neural networks for NMC and graphite structures; users can also train custom networks using GeoDict‑AI for other particle systems.

Key Features

Grain identification: Uses a specialized watershed algorithm to identify individual grains, with grain relinking to correct over‑segmentation; particles on model boundaries can be excluded to avoid statistical bias; ellipsoids, cuboids, or short fibers can be fitted to identified grains.

Statistical data: Generates histograms of diameter, volume, or surface area; 3D visualization of diameter distribution and sphericity; particles can be sorted and individually visualized.

Binder identification (AI): Neural‑network‑based distinction between grains and binders; pre‑trained networks available for graphite anodes (flake‑shaped) and NMC cathodes (spherical).

Typical Applications

Battery materials: statistical analysis of active material particles and binder distribution in electrodes

Digital rock physics: single‑grain information and digital particle size distribution

Filtration: characterization of binders and filter particles

Composites: detection and characterization of unwanted granular impurities

Quality control: detection of binders and impurities in composite materials

GrainFind's particle identification is fully automated, with technical details documented in the user manual for full transparency.

3. PoroDict — Pore Space Characterization

Understanding the pore space of porous materials is critical for optimizing performance in filtration, energy storage, catalysis, and materials science.PoroDict, together with MatDict, provides a powerful solution for comprehensive characterization and analysis of porous media.

PoroDict extracts key pore structure features from CT, µCT, FIB‑SEM data, or GeoDict‑generated structures, enabling precise evaluation of porosity, connectivity, and transport properties.

Key Features

Geometric pore size distribution: Characterizes pore radii by fitting spheres into the pore space (purely geometric,不分 open/closed pores).

Porosimetry: Corresponds to experimental methods like MIP or LEP, calculating the volume of non‑wetting fluid intruded while accounting for connectivity and closed pores.

Pore identification: Segments pore space using watershed algorithms; determines pore count, spatial distribution, sphericity, volume, diameter, orientation, surface area, and contact area.

Open/closed porosity: Calculates volume and count of open pores (connected to the surface) and closed pores (isolated).

Bubble point pressure: Based on the largest through‑pore and the Young‑Laplace equation; provides detailed pore throat analysis.

Percolation paths: Calculates the maximum diameter of spherical particles that can traverse the medium and determines the shortest paths; supports visualization and animation.

Chord length distribution (CLD): Enables precise comparison of porous geometries, especially useful for 2D cross‑sections where direct PSD measurement is difficult.

Geodesic tortuosity: Quantifies the complexity of transport paths by measuring the ratio of shortest actual path to straight‑line distance, helping evaluate permeability and transport efficiency.

PoroDict also includes two powerful automation GeoApps:

Compute Tortuosity: Calculates tortuosity using geometric, physical, and hybrid methods

ASTM Calculation Factor: Simulates bubble point testing per ASTM E3278‑21 for woven wire filter cloth

PoroDict is recommended by ASTM International standards for woven wire filter cloth specifications and testing procedures.

Typical Applications

Battery technology: reveals critical pore structure information affecting energy storage and lifespan

Fuel cells: quantifies transport properties in gas diffusion layers (GDL)

Geoscience: evaluates sandstone geometry for hydrocarbon exploration

Filtration: analyzes pore and transport properties of woven and nonwoven media

4. MatDict — Solid Phase Geometry Analysis

In filtration, energy storage, catalysis, and new material development, the microstructure of a material often determines success or failure.MatDict is a dedicated analysis module that extracts and quantifies key geometric properties of the solid phase, including material thickness, non‑uniformity, particle size distribution, surface area, and connectivity.

Its underlying data can come from high‑resolution 3D models obtained from CT, µCT, FIB‑SEM scans, or GeoDict‑generated structures.The results provide reliable assessments of structural parameters that significantly impact mechanical strength, service life, and functional efficiency.

Key Features

Structural information: Calculates porosity, density, basis weight, and material fraction in all three spatial directions

Thickness estimation: Precise thickness mapping using advanced imaging and computational analysis

2D density maps: Analyzes spatial non‑uniformity by calculating basis weight, solid volume fraction (SVF), and object count distributions on cross‑sections

3D non‑uniformity analysis: Divides the structure into sub‑volumes and computes histograms of solid volume fraction

Solid phase size distribution: Characterizes solid material size by fitting spheres into the solid phase (purely geometric)

Connected regions: Distinguishes separate regions in the microstructure, highlighting connectivity and isolation

Percolation paths: Calculates maximum traversable particle diameter and shortest paths through the solid phase

Surface area estimation: Quantifies the total interfacial area where key physical and chemical processes occur

Three‑phase contact line estimation: Identifies regions where three different phases meet

Minkowski parameters: Quantifies geometric and topological features including volume, surface area, curvature, and connectivity

GAD object orientation: Computes orientation tensors for given object types

Chord length distribution (CLD): Enables precise comparison of complex solid geometries

Geodesic tortuosity: Quantifies transport path complexity through the solid phase

Two‑point correlation function: Provides statistical description of material microstructure

Object analysis: Comprehensive suite for computing properties, distributions, and statistics based on object information

Euclidean distance transform (EDT): Calculates the shortest distance from any point in pore space to the nearest solid boundary

Typical Applications

Battery electrodes: structural insights to improve energy density and lifespan

Fuel cells: analysis of gas diffusion layers to enhance system efficiency

Reservoir analysis: reliable assessment of complex sandstones and reservoir rocks

Filter media: optimization of woven and nonwoven fabrics for high‑performance filtration

MatDict helps you gain deeper insight into your materials, enabling faster and more confident decision‑making — whether designing new materials or optimizing existing ones.